Impact of VI-RADS quality score on VI-RADS diagnostic performance for muscle invasion: the role of tumor size and reader experience.

objectivesTo evaluate the inter-reader agreement of the Vesical Imaging Reporting and Data System (VI-RADS) quality scoring system and to assess its impact on the diagnostic performance for detecting muscle invasion. MATERIALS AND

methodsThis retrospective study included 587 patients who underwent multiparametric bladder MRI before surgery. Three radiologists (two specialists and one non-genitourinary radiologist) independently assessed VI-RADS category and quality scores. Inter-reader agreement was evaluated using weighted kappa statistics. Associations between image quality, VI-RADS categorization, and tumor characteristics were analyzed using multivariable models. Diagnostic performance for muscle invasion was assessed at predefined VI-RADS thresholds and compared across quality groups and tumor size subgroups.

resultsA total of 7/587 (1.2%), 284/587 (48.4%), and 296/587 (50.4%) scans were VI-RADS quality score 0, 1, and 2, respectively. Inter-reader agreement for the final quality score was substantial (k = 0.628-0.702). Higher VI-RADS quality scores were associated with increased T2-weighted image VI-RADS scores (odds ratio [OR], 2.95; 95% confidence interval, 1.17-7.43; p = 0.021). In tumors < 3 cm, particularly for the less experienced reader, higher image quality was associated with increased positive predictive value (PPV) and accuracy at the VI-RADS ≥ 3 threshold (PPV: 51.0% vs 22.7%, p = 0.004; accuracy: 81.6% vs 71.7%, p = 0.023). No significant differences in diagnostic performance were observed across VI-RADS thresholds in tumors ≥ 3 cm.

conclusionThe VI-RADS quality scoring system demonstrated good inter-reader agreement. Optimal image quality improved diagnostic performance in small tumors and for the less experienced reader; however, because only one less experienced reader was included, this finding requires external validation. KEY POINTS: Question Does image quality, assessed by the VI-RADS quality score, influence diagnostic performance of VI-RADS for detecting muscle invasion in bladder cancer? Findings Higher image quality improved PPV and accuracy of VI-RADS, particularly in tumors < 3 cm and for the less experienced reader, while having limited impact in larger tumors. Clinical relevance Ensuring optimal MRI image quality enhances the reliability of VI-RADS interpretation, particularly in small tumors and non-specialized settings, and supports standardized imaging as a prerequisite for accurate bladder cancer staging.